When a user of an electronic device browses or previews video files on the device, a video decoder of the device's operating system (OS) may select multiple frames from each file for decoding and potential display one of the decoded video frames as a thumbnail image of the file in question. If the format of the video file is not natively supported by the OS video decoder, however, a generic thumbnail image may be displayed, which can prevent the user from easily determining the content of the underlying video. Even if the format of the video file is natively supported, the selected frame for the thumbnail image may not be representative of the video content (e.g., shows a meaningless scene or portion of a scene). Moreover, different operating systems may use different algorithms to select the representative frame, which may lead to inconsistent results from the perspective of the user. Additionally, decoding multiple candidate frames for each video file can increase power consumption and create delays in the browsing process. Increased power consumption may be particularly problematic for mobile devices in which battery life is an area of concern.